Доклад Александра посвящен наиболее значимым результатам, которых удалось достигнуть за год. В нем рассмотрены топ-публикаций по RL в 2020м: Learning to summarize from human feedback (): Autonomous navigation of stratospheric balloons using reinforcement learning (Google Brain): Emergent complexity and zero-shot transfer via unsupervised environment design (Google Brain): Asymmetric self-play for automatic goal discovery in robotic manipulation (OpenAI): Never Give Up: Learning Directed Exploration Strategies (DeepMind): Critic Regularized Regression (DeepMind):
Hide player controls
Hide resume playing